Types of Mechanical Puzzles: Slide, Jump, Pack, Cut, Match, Twist

The main types of mechanical puzzles are best sorted by the one action they ask of your hands: sliding pieces, jumping pieces, packing pieces into a frame, cutting a shape and rearranging it, matching colours where pieces meet, and twisting linked layers. Once you know the action, you know the right first move, because each action hides its own kind of trap.

A mechanical puzzle is any puzzle built from physical parts that you move by hand. This guide takes the six actions one at a time, gives a small solved example for most of them, and closes with a comparison table.

What makes a puzzle mechanical?

In a mechanical puzzle the rules live inside the object. Nobody hands you a list of clues. The shape of the tray, the size of each piece and the way the parts touch decide which moves are possible.

That changes the question you ask. Instead of wondering which statement is true, you ask which positions you can reach from here, and which of them bring you closer to the goal. Each type below is a different answer to a simple question: what does a single move look like?

How do sliding puzzles work?

A sliding puzzle is a tray of tiles with one empty gap. The only move is to push a tile that sits next to the gap into it, so the gap wanders around the tray while the tiles shift one step at a time.

Small trays are fully understood: the trays with three, eight and fifteen tiles never need more than 6, 31 and 80 moves ^[1].

Worked example. Take the smallest tray of all: a two-by-two square holding tiles A, B and C plus one gap. There are 24 ways to place three tiles and a gap in four cells, yet only 12 of them can be reached by sliding from the solved position. The farthest of those 12 needs six moves, which agrees with the published figure for this size ^[1].

Why can the other half never be reached? In such a small tray the three tiles can only travel around the ring of four cells, so their order, read clockwise, never changes. An arrangement that reads B, A, C clockwise when the solved tray reads A, B, C cannot be produced by any amount of sliding. Before you spend time on a small sliding puzzle, it is worth checking that the target order can be reached at all.

Step-by-step methods hold up on big trays too. A paper published on 10 July 2015 tested a row-by-row method on 10,000 random trays at every width from 4 to 200, and its measured move counts matched the paper's prediction ^[1].

How to start: think about moving the gap rather than the tiles. Before each tile goes home, walk the gap round to the side the tile must enter from, without disturbing anything you have already finished.

How do jumping puzzles work?

A peg-jumping puzzle is a board of holes, all filled with pegs except one or a few. To move, a peg hops over the peg beside it and lands in the empty hole directly behind, and the hopped peg comes off the board. Most puzzles of this type aim for a lone peg at the end.

The common plus-shaped board has 33 holes ^[2]. With only the centre empty it holds 32 pegs. When the goal is a single peg, every jump removes one peg, so a full solution takes exactly 31 jumps, and you know the count before you begin.

Not every start can be finished. On some 37-hole boards, a full board with only the centre empty has no solution at all ^[2].

Worked example. Take a straight row of five holes with pegs in holes 1, 2 and 4. Only one opening jump is legal: peg 1 jumps over peg 2 into hole 3, leaving pegs in holes 3 and 4. Now there are two ways to finish. Peg 3 can jump over peg 4 into hole 5, or peg 4 can jump back over peg 3 into hole 2. Either way a single peg remains.

How to start: when the jump count is fixed, nothing is lost by planning the ending first. Decide where the last peg should land, clear the outlying arms of the board while there are still pegs nearby to help, and avoid leaving lone pegs that nothing can reach.

How do packing puzzles work?

A packing puzzle hands you a set of pieces and a frame, and asks you to fit every piece inside without overlaps. The catch is that area tells you very little. Pieces whose areas add up perfectly can still refuse to fit, as one well-known set of 24 squares shows: their areas match a 70 by 70 square exactly, yet they cannot fill it ^[3].

Worked example. Try to fit squares with sides 1, 2 and 3 into a 4 by 4 frame. Their areas add up to 14, comfortably less than the 16 cells available. They still do not fit. Wherever the 3 by 3 square goes, the cells left over form a strip only one cell wide along two sides of the frame, and the 2 by 2 square has nowhere to go.

How to start: the big pieces have the fewest places they can go, so put them down first, in corners and along edges. Small pieces are flexible, so keep them back to fill the gaps at the end.

Here is a fitting puzzle where the pieces may turn but never flip:

How do cut-and-rearrange puzzles work?

A dissection puzzle gives you a shape that has been cut into pieces and asks you to rebuild those pieces into a second shape. Nothing is added or lost, so both shapes always have the same area.

Any polygon can be cut up and rebuilt into any other polygon of the same area ^[4], so the real challenge is to manage it with the fewest pieces. A proof published on 27 January 2023 showed that an equilateral triangle cannot become a square using three convex pieces, which means a piece with no dents or inward corners ^[4].

Worked example. Why do the triangle and the square resist each other? Take a triangle with sides of 10. Its area is about 43.30, so the square of equal area has sides of about 6.58 and a diagonal of about 9.31. Even the longest straight line inside that square is shorter than one side of the triangle. The 2023 proof leans on this gap: since the longest line in the square is shorter than a side of the triangle, any three-piece version would need each piece to contain exactly one corner of the triangle ^[4].

How to start: compare the corners and edge lengths of the two shapes. The square's right angles must be built from the pieces, so look for pieces whose corners or straight edges can supply it.

How do colour-matching puzzles work?

In a colour-matching puzzle every piece carries colours, and the goal is to arrange the pieces so that colours agree wherever two pieces touch, or so that each face of the finished shape shows one colour. The trap is orientation: a single piece can sit in many positions, and only one of them is right.

A classic set uses cubes painted with six colours, one colour on each face. If two cubes that can be turned to look the same count as one, there are exactly 30 different cubes ^[5].

Worked example. Here is why there are 30. Stand the cube on one colour. The top can take any of the five colours left. Turn the cube so that one particular side colour faces you; the three remaining side faces can then take the last three colours in six different orders. Five tops times six orders gives 30 ^[5].

The puzzle built on this set is pleasingly strict. Pick any cube, then build a copy twice its size from eight of the other cubes, so that the outside matches the chosen cube and every pair of touching inner faces shares a colour. A paper published on 25 September 2018 showed that such a copy exists for each of the 30 cubes, and that its eight small cubes can always be set out in two distinct ways ^[5].

How to start: in a block built from eight small cubes, every small cube sits at a corner and shows three faces to the outside ^[5]. Settle those outer faces first, because they pin down where each small cube must go, then check the inner faces.

Folding a cube from a flat net trains the same habit of tracking which faces end up side by side:

How do twisting puzzles work?

A twisting puzzle is built from layers joined around a hidden core, so a single turn moves many pieces at once. The same idea appears in cubes, spheres, pyramids and other solids.

Because each turn disturbs pieces you have already placed, it helps to find short move sequences that change a few pieces and return everything else to where it was.

How to start: learn what a sequence does by doing it slowly and then undoing it in reverse order. Watching which pieces move and which stay put teaches you more than memorising long lists.

How do the six types compare?

What does each type reward?

Jumping puzzles reward planning ahead, since a few careless jumps can strand pegs far from each other. Packing and cut-and-rearrange puzzles reward a good eye for shapes. Colour matching and twisting puzzles turn into a search for rules you can reuse.

Whichever you pick, the same habit helps across all six: before each move, ask what it changes and what it leaves alone. A move that undoes earlier work is rarely the right one.

Can you solve a mechanical puzzle on paper?

Yes. Many mechanical puzzles can be drawn, and working them out on paper is a good way to plan before your hands touch the pieces. In the gear riddle below, follow the turn of the big gear through the train of gears to the lanterns: